Aaron K. Baughman

Evolutionary facial feature selection

With the growing number of acquired physiological and behavioral biometric samples, biometric data sets are experiencing tremendous growth. As database sizes increase, exhaustive identification searches by matching with entire biometric feature sets become computationally unmanageable. An evolutionary facial feature selector chooses a set of features from prior contextual or meta face features that reduces the search space. This paper discusses and shows the results of an evolutionary computing approach with agglomerative k-means cluster spaces as input parameters into a LDA evaluation function to select facial features from the Carnegie Mellon University Pose, Illumination, and Expression database of human faces (PIE).

Large scale fingerprint mining

Support Vector Machines (SVM) project feature vectors into a linear or non-linear state space using kernel function(s) and attempts to maximize the margin between classes. The projection of feature vectors into a high dimensional hyperspace structure helps to provide sparse separable clusters of data. Biometric data such as fingerprints transformed into feature vectors are candidates for support vector machine sparse classification. Fingerprint position and ridge flow pattern classification provide a feature vector for a kernel function(s). As samples are projected into a hyperspace construct, fingerprint identification velocity will improve while performance will increase or remain constant. This paper discusses and shows the results of a fingerprint identification support vector machine within a novel hyperspace structure. The support vector machine is formalized with a high dimensional hyperspace structure with an internal bootstrapped c-means clustering algorithm and probabilistic neural network (PNN). The National Institute of Standards and Technology (NIST) provided data set Special Database 14 for all experiments.

IBM Predicts Cloud Computing Demand for Sports Tournaments

The rapid growth of the Internet and of mobile and other smart technologies has generated increased demand on digital platforms, which are supported by enterprise cloud-computing capabilities. To support IBM’s leadership in analytics, mobile, and cloud technologies, a small team within IBM Global Technology Services (GTS) developed a system that uses advanced analytics to address the dynamic and unpredictable Web traffic patterns produced by a digital-enterprise workload, while driving greater operational efficiencies in computing and labor resources. Current cloud platforms are reactive; that is, they require human intervention to scale computing resources to meet demand. To address this shortcoming, the GTS team developed the Predictive Cloud Computing (PCC) system. PCC uses multiple advanced analytical techniques, such as novel numerical analysis techniques, discrete-event simulation, and advanced forecasting to produce models that forecast Internet traffic demands in near real time, allocating computi...

Multimedia Data Mining and Analytics

This chapter gives an overview of multimedia data processing history as a sequence of disruptive innovations and identifies the trends of its future development. Multimedia data processing and mining penetrates into all spheres of human life to improve efficiency of businesses and governments, facilitate social interaction, enhance sporting and entertainment events, and moderate further innovations in science, technology and arts. The disruptive innovations in mobile, social, cognitive, cloud and organic based computing will enable the current and future maturation of multimedia data mining. The chapter concludes with an overview of the other chapters included in the book.

Disruptive Innovation: Large Scale Multimedia Data Mining

This chapter gives an overview of multimedia data processing history as a sequence of Disruptive innovations and identifies the trends of its future development. Multimedia data processing and mining penetrates into all spheres of human life to improve efficiency of businesses and governments, facilitate social interaction, enhance sporting and entertainment events, and moderate further innovations in science, technology and arts. The disruptive innovations in mobile, social, cognitive, cloud and organic based computing will enable the current and future maturation of Multimedia data mining . The chapter concludes with an overview of the other chapters included in the book.

Large-Scale Biometric Multimedia Processing

The field of Biometrics analyses organic signals from people to identify or verify an identity using a combination of physiological, behavioural or cognitive characteristics such as voice, fingerprints, eye color, facial features, iris, handwriting or other characteristics. Large-scale biometric identification systems can benefit from modern optimisation, classification and parallel computation techniques to reduce cost and increase accuracy. This chapter discusses recent and novel developments by the authors in the approaches taken to enable large-scale biometric identification. The authors present an overview of different techniques to perform the tasks of search space reduction, feature selection and parallel processing of biometrics data. Topics covered are: support vector machines and hyperspace transformations for effectively searching extremely large fingerprint databases to identify individuals; evolutionary computing to perform efficient facial feature selection for identification purposes; and cloud and high-performance designs for biometric systems.